The collapse and fragmentation of initially filamentary , magnetic molecular clouds is calculated in three dimensions with a gravitational , radiative hydrodynamics code .
The code includes magnetic field effects in an approximate manner : magnetic pressure , tension , braking , and ambipolar diffusion are all modelled .
The parameters varied are the ratio of the ambipolar diffusion time to the free fall time at the center of the filamentary cloud ( t _ { ad } / t _ { ff } = 10 , 20 , or 10 ^ { 6 } \sim \infty ) , the cloud ’ s reference magnetic field strength ( B _ { oi } = 0 , 200 , or 300 microgauss – the latter two values leading to magnetically subcritical clouds ) , the ratio of rotational to gravitational energy of the filament ( 10 ^ { -4 } or 10 ^ { -2 } ) , and the efficiency of magnetic braking ( represented by a factor f _ { mb } = 0 , 10 ^ { -4 } , or 10 ^ { -3 } ) .
Three types of outcomes are observed : direct collapse and fragmentation into a multiple protostar system ( models with B _ { oi } = 0 ) , periodic contraction and expansion without collapse ( models with t _ { ad } / t _ { ff } = 10 ^ { 6 } ) , or periodic contraction and expansion leading eventually to collapse on a time scale of \sim 6 to 12 t _ { ff } ( all other models ) .
Because the computational grid is a finite volume sphere , the expanding clouds bounce off the spherical boundary and re-collapse toward the center of the spherical grid , leading to the periodic formation of shocked regions where the infalling gas collides with itself , forming dense layers susceptible to sustained collapse and eventual fragmentation .
While the models begin their evolution at rest except for the assumed solid-body rotation , they develop weakly supersonic velocity fields as a result of the rebounding prior to collapse .
The models show that magnetically-supported clouds subject to magnetic braking can undergo dynamic collapse leading to protostellar fragmentation on scales of 10 AU to 100 AU , consistent with typical binary star separations .